Chapter 2 RDF Syntax 1

RDF Overview RDF data model RDF syntax? RDF serializations: XML, Turtle, N3, ntriples RDF Schema (RDFS) Semantics of RDF and RDFS Axiomatic Semantics Operational semantics based on rules Querying RDF via SPARQL

Introduction Problem: What does an XML document mean? XML is about data structures The meaning (semantics) not apparent to machines RDF is more a data model than a language It is realized in many different formats RDF defines very basic semantics RDFS and OWL define more RDF vocabulary for building rich data models RDF remains domain independent

Example 1 What does this mean? <academicStaffMember> Grigoris Antoniou </academicStaffMember> <professor> Michael Maher </professor> <course name="Discrete Mathematics"> <isTaughtBy> David Billington </isTaughtBy> </course> What does this mean? Are professors also academic staff members? If someone teaches a course, are they an academic staff member? Can’t say in XML, but can say so in RDFS

Example 2 Embedding of elements is just a syntactic constraint <course name="Discrete Mathematics"> <lecturer>David Billington</lecturer> </course> <lecturer name="David Billington"> <teaches>Discrete Mathematics</teaches> </lecturer> <teachingOffering> <course>Discrete Mathematics</course> </teachingOffering> Embedding of elements is just a syntactic constraint No meaning is defined Meaning is in documentation or viewer’s minds Does the machine have a mind?

Key RDF documents: standards http://w3.org/standards/techs/rdf

Topics Basic concepts of RDF Vocabularies and modeling Resources, properties, values, statements, triples URIs and URIrefs RDF graphs Literals, qnames Vocabularies and modeling Vocabularies Blank nodes, data modeling, types, reification Lists, bags, collections Serialization of RDF graphs XML, Turtle, Ntriples Critique of RDF

What is RDF? A data model for representing information (esp. metadata) about resources in the Web Can represent information about things that can be identified on the Web, even when not retrievable (e.g., a book) Usecases: provide data for applications rather than directly to people

RDF Basics Core idea: identify resources using Web identifiers and describing resources in terms of simple properties and property values RDF data model is as a “pure” graph model To identify resources, RDF uses Uniform Resource Identifiers (URIs) and URI references (URIrefs). Definition: A resource is anything that is identifiable by a URIref

Example Consider the following information: “there is a Person identified by http://www.w3.org/People/EM/contact#me, whose name is Eric Miller, whose email address is em@w3.org, and whose title is Dr.”

Example (cont’d)

Basics Resources being described have properties that have values, and resources are described by making statements specifing those properties and values The part that identifies the thing the statement is about is the subject The part that identifies the property of the subject the statement specifies is the predicate The part that identifies the property’s value is the object

Example http://www.example.org/index.html has a creator whose value is “John Smith” The subject is the URL http://www.example.org/index.html The predicate is the word "creator" The object is the phrase “John Smith”

RDF Triples RDF statements can be written as triples Simple ntriples notation has a set of triples terminated by a periods, where URI’s are inside angle brackets <http://www.example.org/index.html> <http://purl.org/dc/elements/1.1/creator> <http://www.example.org/staffid/85740> . <http://www.example.org/index.html> <http://www.example.org/terms/creation-date> "August 16, 1999" . <http://www.example.org/index.html> <http://purl.org/dc/elements/1.1/language> "en" .

Graphs: pure and impure

Pure graph model A pure graph model consists only of edges between pairs of nodes Can be directed or undirected; can be labeled or not A graph can be represented as an unordered collection of (subject, predicate, object) triples If directed, predicate goes from subject to object Nodes not the subject or object of a triple are not even allowed (John, likes, Mary), (Mary, likes, Bill), ( John, hates, Bill) john bill mary likes hates

RDF graph model RDF is like this with a few caveats Subjects and predicates are identified by a URI Object can also be a URI but can also be a literals, i.e., a string or a number RDF defines some special URIs and gives them specific meaning http://www.w3.org/1999/02/22-rdf-syntax-ns#type RDF has simple conventions for representing both ordered and unordered sequences and a few other data structures john bill mary likes hates

Property graphs Graph databases have become popular in the past ten years A common extension of the pure graph model is to allow nodes and edges to have properties A simple version: properties are key/value pairs, e.g. Age : 25 Date : 1990-09-21” We might give the spouse edge from John to Mary two properties: start with value “1999-09-1” and end with value “2016-01-11”

URIs and URIREFs

Uniform Resource Identifiers (URIs) URIs identify resources on the Web Unlike URLs, they aren’t limited to identifying things with network locations No organization controls who makes URIs or how they can be used Some URI schemes (http: URL’s) depend on centralized systems such as DNS Others are completely decentralized

URI Reference (URIref) A URIref is a URI with an optional fragment identifier at the end, e.g: http://example.org/index.html#section2 Fragment usecase: HTML fragments refer to a place in a page RDF fragments refer to resources in a RDF graph that the URI denotes, e.g., subjects, predicates or objects http://www.w3.org/2004/02/skos/core : vocabulary for describing topics http://www.w3.org/2004/02/skos/core#broader : the broader concept in SKOS Core vocabulary Like URLs, URIrefs may be either absolute or relative Note: the empty URI refers to the resource it’s in

URIrefs in RDF (cont’d) RDF and Browsers use URIrefs to identify things, but interpret URIrefs slightly differently: Browsers also use URIrefs to retrieve things RDF uses URIrefs only to identify things and these might not even be retrievable Linked Data best practice is to use HTTP URIs that return RDF data for every URI

Content Negotiation What does HTTP stand for?

Content Negotiation What does HTTP stand for? HTTP == Hypertext Transfer Protocol Lets Web client (browser, program) and server (apache) do many things (e.g., authentication) E.g.: specify format of data returned, e.g., as RDF data in one of several formats or as HTML Consider http://dbpedia.org/page/Alan_Turing curl –LI http://dbpedia.org/page/Alan_Turing curl –LH "Accept:application/rdf+xml" http://dbpedia.org/page/Alan_Turing

RDF Graphs

RDF Graphs RDF models statements by nodes and arcs in a graph A statement is represented by a node for the subject, a node for the object and an arc for the predicate (subject => object) A node may be identified by a URIref or it can be a literal or a blank node An arc is identified by a URIref Note: We will draw RDF graphs as directed graphs But an arc can be the subject of an RDF statement :has_parent owl:inverseOf :has_child

Example Consider the following statements: http://www.example.org/index.html has a creation-date whose value is August 16, 1999. http://www.example.org/index.html has a language whose value is English.

The RDF Graph of the Example Note: http://purl.org/dc/elements/1.1 is prefix for the Dublin Core vocabulary/ontology http://www.example.org/… is uses for examples

RDF and Related Data Models In terms of the relational model, an RDF statement is like a tuple in a relation Graph with columns Subject, Predicate, Object For first-order logic, an RDF statement is like an atomic formula triple(subj, pred, obj) where triple is a FOL predicate and subj, pred and obj are constants Alternativedly in logic: pred(subj, obj)

Literals and QNames

What is 27? Number or string? Literals Answer: 27 is a plain literal here !!!!! What is 27? Number or string?

Plain and Typed Literals RDF has two kinds of literals: plain and typed Plain literals have a lexical form (their lexical value) and optionally a language tag, e.g: "27" , "Hello world"@en, "Bonjour le monde" @fr RDF typed literals are formed by pairing a string with a URIref for a particular XMLS datatype, e.g.: “27”^^http://www.w3.org/2001/XMLSchema#integer ”27"^^xsd:int

Data Types for Literals In practice, the most widely used data typing scheme is the one by XML Schema But any externally defined data typing scheme is allowed in RDF documents XML Schema predefines many data types E.g. Booleans, integers, floating-point numbers, times, dates, etc.

XMLSchema Datatypes http://www.w3.org/TR/xmlschema-2/

Qnames for URIrefs The ntriples notation results in very long lines We can use an XML qualified name (QName) w/o brackets for a full URI reference http://dbpedia.org/page/Alan_Turing dbp:Alan_Turing Qnames have a prefix that’s been assigned to a namespace URI, a colon and a local name How to assign a prefix to a URI varies by serialization The concepts of names and namespaces used in RDF originate in XML

Topics Basic concepts of RDF Vocabularies and modeling Resources, properties, values, statements, triples URIs and URIrefs RDF graphs Literals, qnames Vocabularies and modeling Vocabularies Blank nodes, data modeling, types, reification Lists, bags, collections Serialization of RDF graphs XML, Turtle, Ntriples Critique of RDF